/*=============================================================================
# Filename: Util.cpp
# Author: Bookug Lobert
# Mail: 1181955272@qq.com
# Last Modified: 2015-10-16 10:43
# Description:
1. firstly written by liyouhuan, modified by zengli
2. achieve functions in Util.h
=============================================================================*/

#include "Util.h"

using namespace std;

//==================================================================================================================
//configure() to config the basic options of gStore system
//==================================================================================================================

//string Util::profile = "../init.conf";
string Util::profile = "init.conf";

map<string, string> Util::global_config;

//database home directory, which is an absolute path by config
//TODO:everywhere using database, the prefix should be it
//string Util::db_home = ".";

//false:single true:distribute
//bool Util::gstore_mode = false;

//control the debug information
//string Util::debug_level = "simple";

//database placed in which path
//string Util::db_path = ".";

//the suffix to be added to database name
//string Util::db_suffix = ".db";

//the maxium buffer size assigned to gStore system
//string Util::buffer_maxium = "100"; //the unit is GB

//the maxium thread num assigned to gStore system
//string Util::thread_maxium = "1000";

//if record logs in gStore system(to be recoverable or faster)
//string Util::operation_logs = "true";

//==================================================================================================================

//NOTICE:used in Database, Join and Strategy
//int Util::triple_num = 0;
//int Util::pre_num = 0;
//int Util::entity_num = 0;
//int Util::literal_num = 0;

//string Util::tmp_path = "../.tmp/";
//string Util::debug_path = "../.debug/";
string Util::tmp_path = ".tmp/";
string Util::debug_path = ".debug/";

//QUERY: assign all in Util()?
//BETTER:assigned in KVstore, not one tree?
FILE* Util::debug_kvstore = NULL;  //used by KVstore
FILE* Util::debug_database = NULL; //used by Database
FILE* Util::debug_vstree = NULL;   //used by VSTree

string Util::gserver_port_file = "bin/.gserver_port";
string Util::gserver_port_swap = "bin/.gserver_port.swap";
string Util::gserver_log = "logs/gserver.log";

string Util::backup_path = "backups/";

//set hash table
HashFunction Util::hash[] = { Util::simpleHash, Util::APHash, Util::BKDRHash, Util::DJBHash, Util::ELFHash,
                              Util::DEKHash, Util::BPHash, Util::FNVHash, Util::HFLPHash, Util::HFHash, Util::JSHash,
                              Util::PJWHash, Util::RSHash, Util::SDBMHash, Util::StrHash, Util::TianlHash, NULL };

//remove spaces in the left
char*
Util::l_trim(char* szOutput, const char* szInput)
{
  assert(szInput != NULL);
  assert(szOutput != NULL);
  assert(szOutput != szInput);
  for (; *szInput != '\0' && isspace(*szInput); ++szInput)
    ;
  return strcpy(szOutput, szInput);
}

//remove spaces in the right
char*
Util::r_trim(char* szOutput, const char* szInput)
{
  char* p = NULL;
  assert(szInput != NULL);
  assert(szOutput != NULL);
  assert(szOutput != szInput);
  strcpy(szOutput, szInput);
  for (p = szOutput + strlen(szOutput) - 1; p >= szOutput && isspace(*p); --p)
    ;
  *(++p) = '\0';
  return szOutput;
}

//remove spaces in the two sides
char*
Util::a_trim(char* szOutput, const char* szInput)
{
  char* p = NULL;
  assert(szInput != NULL);
  assert(szOutput != NULL);
  l_trim(szOutput, szInput);
  for (p = szOutput + strlen(szOutput) - 1; p >= szOutput && isspace(*p); --p)
    ;
  *(++p) = '\0';
  return szOutput;
}

bool
Util::configure()
{
  const unsigned len = 505;
  char* buf, *c;
  char buf_i[len], buf_o[len];
  FILE* fp = NULL;
  char keyname[len];
  char keyval[len];

  //initialize the settings
  Util::global_config["gstore_mode"] = "single";
  //NOTICE+BETTER+TODO:use macro is better to avoid too many judging on this variable(add a DEBUG macro at the outer)
  Util::global_config["debug_level"] = "simple";
  Util::global_config["db_home"] = ".";
  Util::global_config["db_suffix"] = ".db";
  Util::global_config["buffer_maxium"] = "100";
  Util::global_config["thread_maxium"] = "1000";
  //TODO:to be recoverable
  Util::global_config["operation_logs"] = "true";

#ifdef DEBUG
  fprintf(stderr, "profile: %s\n", profile.c_str());
#endif
  if ((fp = fopen(profile.c_str(), "r")) == NULL) //NOTICE: this is not a binary file
  {
#ifdef DEBUG
    fprintf(stderr, "openfile [%s] error [%s]\n", profile.c_str(), strerror(errno));
#endif
    return false;
  }
  fseek(fp, 0, SEEK_SET);

  while (!feof(fp) && fgets(buf_i, len, fp) != NULL) {
    //fprintf(stderr, "buffer: %s\n", buf_i);
    Util::l_trim(buf_o, buf_i);
    if (strlen(buf_o) <= 0)
      continue;
    buf = NULL;
    buf = buf_o;
    if (buf[0] == '#') {
      continue;
    } else if (buf[0] == '[') {
      continue;
    }
    if ((c = (char*)strchr(buf, '=')) == NULL)
      continue;
    memset(keyname, 0, sizeof(keyname));
    sscanf(buf, "%[^=|^ |^\t]", keyname);
#ifdef DEBUG
//fprintf(stderr, "keyname: %s\n", keyname);
#endif
    sscanf(++c, "%[^\n]", keyval);
    char* keyval_o = (char*)calloc(strlen(keyval) + 1, sizeof(char));
    if (keyval_o != NULL) {
      Util::a_trim(keyval_o, keyval);
#ifdef DEBUG
//fprintf(stderr, "keyval: %s\n", keyval_o);
#endif
      if (keyval_o && strlen(keyval_o) > 0) {
        //strcpy(keyval, keyval_o);
        global_config[string(keyname)] = string(keyval_o);
      }
      xfree(keyval_o);
    }
  }

  fclose(fp);
  //display all settings here
  cout << "the current settings are as below: " << endl;
  cout << "key : value" << endl;
  cout << "------------------------------------------------------------" << endl;
  for (map<string, string>::iterator it = global_config.begin(); it != global_config.end(); ++it) {
    cout << it->first << " : " << it->second << endl;
  }
  cout << endl;

  return true;
  //return Util::config_setting() && Util::config_debug() && Util::config_advanced();
}

bool
Util::config_debug()
{
  const unsigned len1 = 100;
  const unsigned len2 = 505;
  char AppName[] = "setting";
  char KeyName[] = "mode";
  char appname[len1], keyname[len1];
  char KeyVal[len1];
  char* buf, *c;
  char buf_i[len1], buf_o[len1];
  FILE* fp = NULL;
  int status = 0; // 1 AppName 2 KeyName
  return true;
}

bool
Util::config_advanced()
{
  const unsigned len1 = 100;
  const unsigned len2 = 505;
  char AppName[] = "setting";
  char KeyName[] = "mode";
  char appname[len1], keyname[len1];
  char KeyVal[len1];
  char* buf, *c;
  char buf_i[len1], buf_o[len1];
  FILE* fp = NULL;
  int status = 0; // 1 AppName 2 KeyName
  return true;
}

bool
Util::config_setting()
{
  const unsigned len1 = 100;
  const unsigned len2 = 505;
  char AppName[] = "setting";
  char KeyName[] = "mode";
  char appname[len1], keyname[len1];
  char KeyVal[len1];
  char* buf, *c;
  char buf_i[len2], buf_o[len2];
  FILE* fp = NULL;
  int status = 0; // 1 AppName 2 KeyName

#ifdef DEBUG
  fprintf(stderr, "profile: %s\n", profile.c_str());
#endif
  if ((fp = fopen(profile.c_str(), "r")) == NULL) //NOTICE: this is not a binary file
  {
#ifdef DEBUG
    fprintf(stderr, "openfile [%s] error [%s]\n", profile.c_str(), strerror(errno));
#endif
    return false;
  }
  fseek(fp, 0, SEEK_SET);
  memset(appname, 0, sizeof(appname));
  sprintf(appname, "[%s]", AppName);
#ifdef DEBUG
  fprintf(stderr, "appname: %s\n", appname);
#endif

  while (!feof(fp) && fgets(buf_i, len2, fp) != NULL) {
    //fprintf(stderr, "buffer: %s\n", buf_i);
    Util::l_trim(buf_o, buf_i);
    if (strlen(buf_o) <= 0)
      continue;
    buf = NULL;
    buf = buf_o;
    if (buf[0] == '#') {
      continue;
    }
    if (status == 0) {
      if (strncmp(buf, appname, strlen(appname)) == 0) {
#ifdef DEBUG
        fprintf(stderr, "app found!\n");
#endif
        status = 1;
        continue;
      }
    } else if (status == 1) {
      if (buf[0] == '[') {
        //NOTICE: nested module is not allowed
        break;
      } else {
        if ((c = (char*)strchr(buf, '=')) == NULL)
          continue;
        memset(keyname, 0, sizeof(keyname));
        sscanf(buf, "%[^=|^ |^\t]", keyname);
#ifdef DEBUG
        fprintf(stderr, "keyname: %s\n", keyname);
#endif
        if (strcmp(keyname, KeyName) == 0) {
#ifdef DEBUG
          fprintf(stderr, "key found!\n");
#endif
          sscanf(++c, "%[^\n]", KeyVal);
          char* KeyVal_o = (char*)calloc(strlen(KeyVal) + 1, sizeof(char));
          if (KeyVal_o != NULL) {
            Util::a_trim(KeyVal_o, KeyVal);
#ifdef DEBUG
            fprintf(stderr, "KeyVal: %s\n", KeyVal_o);
#endif
            if (KeyVal_o && strlen(KeyVal_o) > 0)
              strcpy(KeyVal, KeyVal_o);
            xfree(KeyVal_o);
          }
          status = 2;
          break;
        }
      }
    }
  }
  fclose(fp);
  //if(found == 2)
  //return(0);
  //else
  //return(-1);
  //fprintf(stderr, "%s\n", KeyVal);
  if (strcmp(KeyVal, "distribute") == 0) {
#ifdef DEBUG
    fprintf(stderr, "the gStore will run in distributed mode!\n");
#endif
    //Util::gStore_mode = true;
  }

  return true; //config success
}

Util::Util()
{
  Util::configure();
#ifdef DEBUG_KVSTORE
  if (this->debug_kvstore == NULL) {
    string s = this->debug_path + "kv.log";
    this->debug_kvstore = fopen(s.c_str(), "w+");
    if (this->debug_kvstore == NULL) {
      cerr << "open error: kv.log\n";
      this->debug_kvstore = stderr;
    }
  }
#endif
#ifdef DEBUG_DATABASE
  if (this->debug_database == NULL) {
    string s = this->debug_path + "db.log";
    this->debug_database = fopen(s.c_str(), "w+");
    if (this->debug_database == NULL) {
      cerr << "open error: db.log\n";
      this->debug_database = stderr;
    }
  }
#endif
#ifdef DEBUG_VSTREE
  if (this->debug_vstree == NULL) {
    string s = this->debug_path + "vs.log";
    this->debug_vstree = fopen(s.c_str(), "w+");
    if (this->debug_vstree == NULL) {
      cerr << "open error: vs.log\n";
      this->debug_vstree = stderr;
    }
  }
#endif
}

Util::~Util()
{
#ifdef DEBUG_KVSTORE
  fclose(this->debug_kvstore); //NULL is ok, just like free(NULL)
  this->debug_kvstore = NULL;
#endif
#ifdef DEBUG_DATABASE
  fclose(this->debug_database); //NULL is ok, just like free(NULL)
  this->debug_database = NULL;
#endif
}

string
Util::getThreadID()
{
  //thread::id, neither int or long
  auto myid = this_thread::get_id();
  stringstream ss;
  ss << myid;
  return ss.str();
}

int
Util::memUsedPercentage()
{
  FILE* fp = fopen("/proc/meminfo", "r");
  if (fp == NULL)
    return -1;
  char str[20], tail[3];
  unsigned t, sum, used = 0;              //WARN:unsigned,memory cant be too large!
  fscanf(fp, "%s%u%s", str, &sum, tail);  //MemTotal, KB
  fscanf(fp, "%s%u%s", str, &used, tail); //MemFree
  fscanf(fp, "%s%u%s", str, &t, tail);
  if (strcmp(str, "MemAvailable") == 0) {
    //QUERY:what is the relation between MemFree and MemAvailable?
    used = t;
    //scanf("%s%u%s", str, &t, tail);		//Buffers
    //used += t;
    //scanf("%s%u%s", str, &t, tail);		//Cached
    //used += t;
  }
  //else							//Buffers
  //{
  //	scanf("%s%u%s", str, &t, tail);		//Cached
  //	used += t;
  //}
  used = sum - used;
  fclose(fp);
  return (int)(used * 100.0 / sum);
}

int
Util::memoryLeft()
{
  FILE* fp = fopen("/proc/meminfo", "r");
  if (fp == NULL)
    return -1;

  char str[20], tail[3];
  unsigned num, avail = 0, free = 0, buffer = 0, cache = 0; //WARN:unsigned,memory cant be too large!
  while (fscanf(fp, "%s%u%s", str, &num, tail) != EOF) {
    if (strcmp(str, "MemAvailable:") == 0)
      avail = num;
    if (strcmp(str, "MemFree:") == 0)
      free = num;
    if (strcmp(str, "Buffers:") == 0)
      buffer = num;
    if (strcmp(str, "Cached:") == 0)
      cache = num;
  }

  if (avail == 0)
    avail = free + buffer + cache;

  fclose(fp);
  return avail / Util::MB;
}

bool
Util::is_literal_ele(TYPE_ENTITY_LITERAL_ID _id)
{
  return _id >= Util::LITERAL_FIRST_ID;
}

bool
Util::is_entity_ele(TYPE_ENTITY_LITERAL_ID id)
{
  return id < Util::LITERAL_FIRST_ID;
}

bool
Util::isEntity(const std::string& _str)
{
  if (_str[0] == '<') {
    return true;
  } else {
    return false;
  }
}

bool
Util::isLiteral(const std::string& _str)
{
  if (_str[0] == '"') {
    return true;
  } else {
    return false;
  }
}

//NOTICE: require that the list is ordered
unsigned
Util::removeDuplicate(unsigned* _list, unsigned _len)
{
  if (_list == NULL || _len == 0) {
    return 0;
  }
  unsigned valid = 0, limit = _len - 1;
  for (unsigned i = 0; i < limit; ++i) {
    if (_list[i] != _list[i + 1]) {
      _list[valid++] = _list[i];
    }
  }
  _list[valid++] = _list[limit];

  return valid;
}

int
Util::cmp_int(const void* _i1, const void* _i2)
{
  return *(int*)_i1 - *(int*)_i2;
}

int
Util::cmp_unsigned(const void* _i1, const void* _i2)
{
  unsigned t1 = *(unsigned*)_i1;
  unsigned t2 = *(unsigned*)_i2;
  if (t1 > t2) {
    return 1;
  } else if (t1 == t2) {
    return 0;
  } else //t1 < t2
  {
    return -1;
  }
}

bool
Util::parallel_cmp_unsigned(unsigned _i1, unsigned _i2)
{
  return _i1 < _i2;
}

void
Util::sort(unsigned*& _id_list, unsigned _list_len)
{
#ifndef PARALLEL_SORT
  qsort(_id_list, _list_len, sizeof(unsigned), Util::cmp_unsigned);
#else
  omp_set_num_threads(thread_num);
  __gnu_parallel::sort(_id_list, _id_list + _list_len, Util::parallel_cmp_unsigned);
#endif
}

unsigned
Util::bsearch_int_uporder(unsigned _key, const unsigned* _array, unsigned _array_num)
{
  if (_array_num == 0) {
    //return -1;
    return INVALID;
  }
  if (_array == NULL) {
    //return -1;
    return INVALID;
  }

  unsigned _first = _array[0];
  unsigned _last = _array[_array_num - 1];

  if (_last == _key) {
    return _array_num - 1;
  }

  if (_last < _key || _first > _key) {
    //return -1;
    return INVALID;
  }

  unsigned low = 0;
  unsigned high = _array_num - 1;

  unsigned mid;
  while (low <= high) {
    mid = (high - low) / 2 + low;
    if (_array[mid] == _key) {
      return mid;
    }
    if (_array[mid] > _key) {
      high = mid - 1;
    } else {
      low = mid + 1;
    }
  }

  //return -1;
  return INVALID;
}

bool
Util::bsearch_preid_uporder(TYPE_PREDICATE_ID _preid, unsigned* _pair_idlist, unsigned _list_len)
{
  if (_list_len == 0) {
    return false;
  }
  //NOTICE: if list len > 0, then it must >= 2, so pair num >= 1
  unsigned pair_num = _list_len / 2;
  unsigned _first = _pair_idlist[2 * 0 + 0];
  unsigned _last = _pair_idlist[2 * (pair_num - 1) + 0];

  if (_preid == _last) {
    return true;
  }

  bool not_find = (_last < _preid || _first > _preid);
  if (not_find) {
    return false;
  }

  unsigned low = 0;
  unsigned high = pair_num - 1;
  unsigned mid;

  while (low <= high) {
    mid = (high - low) / 2 + low;
    if (_pair_idlist[2 * mid + 0] == _preid) {
      return true;
    }

    if (_pair_idlist[2 * mid + 0] > _preid) {
      high = mid - 1;
    } else {
      low = mid + 1;
    }
  }

  return false;
}

unsigned
Util::bsearch_vec_uporder(unsigned _key, const vector<unsigned>* _vec)
{
  unsigned tmp_size = _vec->size();
  if (tmp_size == 0) {
    //return -1;
    return INVALID;
  }

  unsigned _first = (*_vec)[0];
  unsigned _last = (*_vec)[tmp_size - 1];

  if (_key == _last) {
    return tmp_size - 1;
  }

  bool not_find = (_last < _key || _first > _key);
  if (not_find) {
    //return -1;
    return INVALID;
  }

  unsigned low = 0;
  unsigned high = tmp_size - 1;
  unsigned mid;

  while (low <= high) {
    mid = (high - low) / 2 + low;
    if ((*_vec)[mid] == _key) {
      return mid;
    }

    if ((*_vec)[mid] > _key) {
      high = mid - 1;
    } else {
      low = mid + 1;
    }
  }

  //return -1;
  return INVALID;
}

string
Util::result_id_str(vector<unsigned*>& _v, int _var_num)
{
  stringstream _ss;

  unsigned size = _v.size();
  for (unsigned i = 0; i < size; ++i) {
    unsigned* _p_int = _v[i];
    _ss << "[";
    for (int j = 0; j < _var_num - 1; j++) {
      _ss << _p_int[j] << ",";
    }
    _ss << _p_int[_var_num - 1] << "]\t";
  }

  return _ss.str();
}

bool
Util::dir_exist(const string _dir)
{
  DIR* dirptr = opendir(_dir.c_str());
  if (dirptr != NULL) {
    closedir(dirptr);
    return true;
  }

  return false;
}

bool
Util::create_dir(const string _dir)
{
  if (!Util::dir_exist(_dir)) {
    mkdir(_dir.c_str(), 0755);
    return true;
  }

  return false;
}

bool
Util::create_file(const string _file)
{
  if (creat(_file.c_str(), 0755) > 0) {
    return true;
  }
  return false;
}

long
Util::get_cur_time()
{
  timeval tv;
  gettimeofday(&tv, NULL);
  return (tv.tv_sec * 1000 + tv.tv_usec / 1000);
}

string
Util::get_date_time()
{
  time_t timep;
  time(&timep);
  char tmp[64];
  strftime(tmp, sizeof(tmp), "%Y-%m-%d %H:%M:%S", localtime(&timep));
  return tmp;
}

string
Util::get_timestamp()
{
  string timestamp;
  time_t timep;
  time(&timep);
  char year[5];
  char tmp[64];
  strftime(year, sizeof(year), "%Y", localtime(&timep));
  timestamp += year[2];
  timestamp += year[3];
  strftime(tmp, sizeof(tmp), "%m%d%H%M%S", localtime(&timep));
  timestamp += tmp;
  return timestamp;
}

string
Util::get_backup_time(const string path, const string db_name)
{
  string _db_name = db_name + ".db";
  string::size_type position;
  position = path.find(_db_name);

  if (position == path.npos)
    return "";
  string db_folder = path.substr(position, path.length());
  string timestamp = db_folder.substr(_db_name.length() + 1, db_folder.length());
  string year, month, day, hour, minute, second;
  year = "20" + timestamp.substr(0, 2);
  month = timestamp.substr(2, 2);
  day = timestamp.substr(4, 2);
  hour = timestamp.substr(6, 2);
  minute = timestamp.substr(8, 2);
  second = timestamp.substr(10, 2);
  string time = year + '-' + month + '-' + day + ' ' + hour + ":" + minute + ":" + second;
  return time;
}

string
Util::get_folder_name(const string path, const string db_name)
{
  string _db_name = db_name + ".db";
  string::size_type position;
  position = path.find(_db_name);

  string db_folder = path.substr(position, path.length());
  return db_folder;
}

bool
Util::save_to_file(const char* _dir, const string _content)
{
  ofstream fout(_dir);

  if (fout.is_open()) {
    fout << _content;
    fout.close();
  }

  return false;
}

int
Util::compare(const char* _str1, unsigned _len1, const char* _str2, unsigned _len2)
{
  int ifswap = 1; //1 indicate: not swapped
  if (_len1 > _len2) {
    const char* str = _str1;
    _str1 = _str2;
    _str2 = str;
    unsigned len = _len1;
    _len1 = _len2;
    _len2 = len;
    ifswap = -1;
  }
  unsigned i;
  //DEBUG: if char can be negative, which cause problem when comparing(128+)
  //
  //NOTICE:little-endian-storage, when string buffer poniter is changed to
  //unsigned long long*, the first char is the lowest byte!
  /*
       unsigned long long *p1 = (unsigned long long*)_str1, *p2 = (unsigned long long*)_str2;
       unsigned limit = _len1/8;
       for(i = 0; i < limit; ++i, ++p1, ++p2)
       {
       if((*p1 ^ *p2) == 0)	continue;
       else
       {
       if(*p1 < *p2)	return -1 * ifswap;
       else			return 1 * ifswap;
       }
       }
       for(i = 8 * limit; i < _len1; ++i)
       {
       if(_str1[i] < _str2[i])	return -1 * ifswap;
       else if(_str1[i] > _str2[i])	return 1 * ifswap;
       else continue;
       }
       if(i == _len2)	return 0;
       else	return -1 * ifswap;
       */
  for (i = 0; i < _len1; ++i) { //ASCII: 0~127 but c: 0~255(-1) all transfered to unsigned char when comparing
    if ((unsigned char)_str1[i] < (unsigned char)_str2[i])
      return -1 * ifswap;
    else if ((unsigned char)_str1[i] > (unsigned char)_str2[i])
      return 1 * ifswap;
    else
      ;
  }
  if (i == _len2)
    return 0;
  else
    return -1 * ifswap;
}

int
Util::string2int(string s)
{
  return atoi(s.c_str());
}

string
Util::int2string(long n)
{
  string s;
  stringstream ss;
  ss << n;
  ss >> s;
  return s;
}

//NOTICE: there does not exist itoa() function in Linux, atoi() is included in stdlib.h
//itoa() is not a standard C function, and it is only used in Windows.
//However, there do exist a function called sprintf() in standard library which can replace itoa()
//char str[255];
//sprintf(str, "%x", 100); //将100转为16进制表示的字符串
char*
Util::itoa(int num, char* str, int radix) //the last parameter means the number's radix: decimal, or octal formats
{
  //index table
  char index[] = "0123456789ABCDEF";
  unsigned unum;
  int i = 0, j, k;
  if (radix == 10 && num < 0) //negative in decimal
  {
    unum = (unsigned)-num;
    str[i++] = '-';
  } else
    unum = (unsigned)num;
  do {
    str[i++] = index[unum % (unsigned)radix];
    unum /= radix;
  } while (unum);
  str[i] = '\0';
  //reverse order
  if (str[0] == '-')
    k = 1;
  else
    k = 0;
  char temp;
  for (j = k; j <= (i - 1) / 2; j++) {
    temp = str[j];
    str[j] = str[i - 1 + k - j];
    str[i - 1 + k - j] = temp;
  }
  return str;
}

string
Util::showtime()
{
  //	fputs("\n\n", logsfp);
  time_t now;
  time(&now);
  //	fputs(ctime(&now), logsfp);
  return string("\n\n") + ctime(&now);
}

string
Util::getQueryFromFile(const char* _file_path)
{
#ifdef DEBUG_PRECISE
  cerr << "file to open: " << _file_path << endl;
#endif
  char buf[10000];
  std::string query_file;

  ifstream fin(_file_path);
  if (!fin) {
    printf("can not open: %s\n", _file_path);
    return "";
  }

  memset(buf, 0, sizeof(buf));
  stringstream _ss;
  while (!fin.eof()) {
    fin.getline(buf, 9999);
    _ss << buf << "\n";
  }
  fin.close();

  return _ss.str();
}

string
Util::getItemsFromDir(string _path)
{
  DIR* dp = NULL;
  struct dirent* entry;
  string ret = "";
  if ((dp = opendir(_path.c_str())) == NULL) {
    fprintf(stderr, "error opening directory!\n");
  } else {
    while ((entry = readdir(dp)) != NULL) {
#ifdef DEBUG_PRECISE
      fprintf(stderr, "%s\n", entry->d_name);
#endif
      string name = string(entry->d_name);
      int len = name.length();
      if (len <= 3) {
        continue;
      }

      //BETTER: change to use db_suffix instead
      if (name.substr(len - 3, 3) == ".db") {
        string tmp = name.substr(0, len - 3);
        if (ret == "")
          ret = tmp;
        else
          ret = ret + "  " + tmp;
      }
    }
    closedir(dp);
  }
#ifdef DEBUG_PRECISE
  fprintf(stderr, "OUT getItemsFromDir\n");
#endif
  return ret;
}

//NOTICE: system() is implemented by fork() and exec(), the latter will change the whole control flow,
//so it must be used with fork()
//The fork() will copy the whole process's image, including the heap and stack.
//But in UNIX, this copy is only logical, in practice it is copy-on-write, which will save a lot memory usage.
//Another function is vfork(), which won't copy a lot of things
//
//http://www.cnblogs.com/wuchanming/p/3784862.html
//http://www.cnblogs.com/sky-heaven/p/4687489.html
//
//BETTER: sysytem() may have some risks, using popen() instead?
string
Util::getSystemOutput(string cmd)
{
  string ans = "";
  string file = Util::tmp_path;
  file += "ans.txt";
  cmd += " > ";
  cmd += file;
  //cerr << cmd << endl;
  int ret = system(cmd.c_str());
  cmd = "rm -rf " + file;
  if (ret < 0) {
    fprintf(stderr, "system call failed!\n");
    system(cmd.c_str());
    return NULL;
  }

  ifstream fin(file.c_str());
  if (!fin) {
    cerr << "getSystemOutput: Fail to open : " << file << endl;
    return NULL;
  }

  string temp;
  getline(fin, temp);
  while (!fin.eof()) {
    //cout<<"system line"<<endl;
    if (ans == "")
      ans = temp;
    else
      ans = ans + "\n" + temp;
    getline(fin, temp);
  }
  fin.close();
  //FILE *fp = NULL;
  //if((fp = fopen(file.c_str(), "r")) == NULL)
  //{
  //fprintf(stderr, "unbale to open file: %s\n", file.c_str());
  //}
  //else
  //{
  //char *ans = (char *)malloc(100);
  //fgets(path, 100, fp);
  //char *find = strchr(path, '\n');
  //if(find != NULL)
  //*find = '\0';
  //fclose(fp);
  //}
  system(cmd.c_str());
  //cerr<<"ans: "<<ans<<endl;
  return ans;
}

// Get the exact file path from given string: ~, ., symbol links
string
Util::getExactPath(const char* str)
{
  string cmd = "realpath ";
  cmd += string(str);

  return getSystemOutput(cmd);
}

void
Util::logging(string _str)
{
  _str += "\n";
#ifdef DEBUG_DATABASE
  fputs(_str.c_str(), Util::debug_database);
  fflush(Util::debug_database);
#endif

#ifdef DEBUG_VSTREE
  fputs(_str.c_str(), Util::debug_vstree);
  fflush(Util::debug_vstree);
#endif
}

unsigned
Util::BKDRHash(const char* _str)
{
  unsigned int seed = 131; // 31 131 1313 13131 131313 etc..
  unsigned int key = 0;

  //for(unsigned i = 0; i < i; ++i)
  while (*_str) {
    //key = key * seed + _str[i];
    key = key * seed + *(_str++);
  }

  return (key & 0x7FFFFFFF);
}

unsigned
Util::simpleHash(const char* _str)
{
  unsigned int key;
  unsigned char* p;

  for (key = 0, p = (unsigned char*)_str; *p; p++)
    key = 31 * key + *p;

  return (key & 0x7FFFFFFF);
}

unsigned
Util::RSHash(const char* _str)
{
  unsigned int b = 378551;
  unsigned int a = 63689;
  unsigned int key = 0;

  while (*_str) {
    key = key * a + (*_str++);
    a *= b;
  }

  return (key & 0x7FFFFFFF);
}

unsigned
Util::JSHash(const char* _str)
{
  unsigned int key = 1315423911;

  while (*_str) {
    key ^= ((key << 5) + (*_str++) + (key >> 2));
  }

  return (key & 0x7FFFFFFF);
}

unsigned
Util::PJWHash(const char* _str)
{
  unsigned int bits_in_unsigned_int = (unsigned int)(sizeof(unsigned int) * 8);
  unsigned int three_quarters = (unsigned int)((bits_in_unsigned_int * 3) / 4);
  unsigned int one_eighth = (unsigned int)(bits_in_unsigned_int / 8);

  unsigned int high_bits = (unsigned int)(0xFFFFFFFF) << (bits_in_unsigned_int - one_eighth);
  unsigned int key = 0;
  unsigned int test = 0;

  while (*_str) {
    key = (key << one_eighth) + (*_str++);
    if ((test = key & high_bits) != 0) {
      key = ((key ^ (test >> three_quarters)) & (~high_bits));
    }
  }

  return (key & 0x7FFFFFFF);
}

unsigned
Util::ELFHash(const char* _str)
{
  unsigned int key = 0;
  unsigned int x = 0;

  while (*_str) {
    key = (key << 4) + (*_str++);
    if ((x = key & 0xF0000000L) != 0) {
      key ^= (x >> 24);
      key &= ~x;
    }
  }

  return (key & 0x7FFFFFFF);
}

unsigned
Util::SDBMHash(const char* _str)
{
  unsigned int key = 0;

  while (*_str) {
    key = (*_str++) + (key << 6) + (key << 16) - key;
  }

  return (key & 0x7FFFFFFF);
}

unsigned
Util::DJBHash(const char* _str)
{
  unsigned int key = 5381;
  while (*_str) {
    key += (key << 5) + (*_str++);
  }
  return (key & 0x7FFFFFFF);
}

unsigned
Util::APHash(const char* _str)
{
  unsigned int key = 0;
  int i;

  for (i = 0; *_str; i++) {
    if ((i & 1) == 0) {
      key ^= ((key << 7) ^ (*_str++) ^ (key >> 3));
    } else {
      key ^= (~((key << 11) ^ (*_str++) ^ (key >> 5)));
    }
  }

  return (key & 0x7FFFFFFF);
}

unsigned
Util::DEKHash(const char* _str)
{
  unsigned int hash = strlen(_str);
  for (; *_str; _str++) {
    hash = ((hash << 5) ^ (hash >> 27)) ^ (*_str);
  }
  return hash;
}

unsigned
Util::BPHash(const char* _str)
{
  unsigned int hash = 0;
  for (; *_str; _str++) {
    hash = hash << 7 ^ (*_str);
  }

  return hash;
}

unsigned
Util::FNVHash(const char* _str)
{
  const unsigned int fnv_prime = 0x811C9DC5;
  unsigned int hash = 0;

  for (; *_str; _str++) {
    hash *= fnv_prime;
    hash ^= (*_str);
  }

  return hash;
}

unsigned
Util::HFLPHash(const char* _str)
{
  unsigned int n = 0;
  char* b = (char*)&n;
  unsigned int len = strlen(_str);
  for (unsigned i = 0; i < len; ++i) {
    b[i % 4] ^= _str[i];
  }
  return n % len;
}

unsigned
Util::HFHash(const char* _str)
{
  int result = 0;
  const char* ptr = _str;
  int c;
  unsigned int len = strlen(_str);
  for (int i = 1; (c = *ptr++); i++)
    result += c * 3 * i;
  if (result < 0)
    result = -result;
  return result % len;
}

unsigned
Util::StrHash(const char* _str)
{
  register unsigned int h;
  register unsigned char* p;
  for (h = 0, p = (unsigned char*)_str; *p; p++) {
    h = 31 * h + *p;
  }

  return h;
}

unsigned
Util::TianlHash(const char* _str)
{
  unsigned long urlHashValue = 0;
  int ilength = strlen(_str);
  int i;
  unsigned char ucChar;
  if (!ilength) {
    return 0;
  }
  if (ilength <= 256) {
    urlHashValue = 16777216 * (ilength - 1);
  } else {
    urlHashValue = 42781900080;
  }
  if (ilength <= 96) {
    for (i = 1; i <= ilength; i++) {
      ucChar = _str[i - 1];
      if (ucChar <= 'Z' && ucChar >= 'A') {
        ucChar = ucChar + 32;
      }
      urlHashValue += (3 * i * ucChar * ucChar + 5 * i * ucChar + 7 * i + 11 * ucChar) % 1677216;
    }
  } else {
    for (i = 1; i <= 96; i++) {
      ucChar = _str[i + ilength - 96 - 1];
      if (ucChar <= 'Z' && ucChar >= 'A') {
        ucChar = ucChar + 32;
      }
      urlHashValue += (3 * i * ucChar * ucChar + 5 * i * ucChar + 7 * i + 11 * ucChar) % 1677216;
    }
  }

  return urlHashValue;
}

//NOTICE: the time of log() and sqrt() in C can be seen as constant

//NOTICE:_b must >= 1
double
Util::logarithm(double _a, double _b)
{
  //REFRENCE: http://blog.csdn.net/liyuanbhu/article/details/8997850
  //a>0 != 1; b>0 (b>=2 using log/log10/change, 1<b<2 using log1p, b<=1?)
  if (_a <= 1 || _b < 1)
    return -1.0;
  double under = log(_a);
  if (_b == 1)
    return 0.0;
  else if (_b < 2)
    return log1p(_b - 1) / under;
  else //_b >= 2
    return log(_b) / under;
  return -1.0;
}

void
Util::intersect(unsigned*& _id_list, unsigned& _id_list_len, const unsigned* _list1, unsigned _len1, const unsigned* _list2, unsigned _len2)
{
  vector<unsigned> res;
  //cout<<"intersect prevar: "<<_len1<<"   "<<_len2<<endl;
  if (_list1 == NULL || _len1 == 0 || _list2 == NULL || _len2 == 0) {
    _id_list = NULL;
    _id_list_len = 0;
  }

  //when size is almost the same, intersect O(n)
  //when one size is small ratio, search in the larger one O(mlogn)
  //
  //n>0 m=nk(0<k<1)
  //compare n(k+1) and nklogn: k0 = log(n/2)2 requiring that n>2
  //k<=k0 binary search; k>k0 intersect
  int method = -1; //0: intersect 1: search in list1 2: search in list2
  unsigned n = _len1;
  double k = 0;
  if (n < _len2) {
    k = (double)n / (double)_len2;
    n = _len2;
    method = 2;
  } else {
    k = (double)_len2 / (double)n;
    method = 1;
  }
  if (n <= 2)
    method = 0;
  else {
    double limit = Util::logarithm(n / 2, 2);
    if (k > limit)
      method = 0;
  }

  switch (method) {
  case 0: { //this bracket is needed if vars are defined in case
    unsigned id_i = 0;
    unsigned num = _len1;
    for (unsigned i = 0; i < num; ++i) {
      unsigned can_id = _list1[i];
      while ((id_i < _len2) && (_list2[id_i] < can_id)) {
        id_i++;
      }

      if (id_i == _len2) {
        break;
      }

      if (can_id == _list2[id_i]) {
        res.push_back(can_id);
        id_i++;
      }
    }
    break;
  }
  case 1: {
    for (unsigned i = 0; i < _len2; ++i) {
      if (Util::bsearch_int_uporder(_list2[i], _list1, _len1) != INVALID)
        res.push_back(_list2[i]);
    }
    break;
  }
  case 2: {
    unsigned m = _len1, i;
    for (i = 0; i < m; ++i) {
      unsigned t = _list1[i];
      if (Util::bsearch_int_uporder(t, _list2, _len2) != INVALID)
        res.push_back(t);
    }
    break;
  }
  default:
    cerr << "no such method in Util::intersect()" << endl;
    break;
  }

  _id_list_len = res.size();

  if (_id_list_len == 0) {
    _id_list = NULL;
  } else {
    _id_list = new unsigned[_id_list_len];
    for (unsigned i = 0; i < _id_list_len; ++i)
      _id_list[i] = res[i];
  }
  delete[] _list1;
  delete[] _list2;
}

int
Util::compIIpair(int _a1, int _b1, int _a2, int _b2)
{
  if (_a1 == _a2 && _b1 == _b2)
    return 0;
  else if (_a1 < _a2 || (_a1 == _a2 && _b1 <= _b2))
    return -1;
  else
    return 1;
}

bool
Util::isValidPort(string str)
{
  //valid port number: 0 - 65535
  if (str.length() < 1 || str.length() > 5) {
    return false;
  }

  unsigned i;
  for (i = 0; i < str.length(); i++) {
    if (str[i] < '0' || str[i] > '9') {
      return false;
    }
  }

  int port = Util::string2int(str);
  if (port < 0 || port > 65535) {
    return false;
  }

  return true;
}

bool
Util::isValidIP(string str)
{
  if (str == "localhost") {
    return true;
  }
  return (Util::isValidIPV4(str) || Util::isValidIPV6(str));
}

bool
Util::isValidIPV4(string str)
{
  regex_t reg;
  char pattern[] = "^(([01]?[0-9][0-9]?|2[0-4][0-9]|25[0-5])\\.){3}([01]?[0-9][0-9]?|2[0-4][0-9]|25[0-5])$";
  regcomp(&reg, pattern, REG_EXTENDED | REG_NOSUB);
  regmatch_t pm[1];
  int status = regexec(&reg, str.c_str(), 1, pm, 0);
  regfree(&reg);
  if (status == REG_NOMATCH) {
    return false;
  }
  return true;
}

bool
Util::isValidIPV6(string str)
{
  //TO BE IMPLEMENTED
  return false;
}

string
Util::getTimeName()
{
  //NOTICE: this is another method to get the concrete time
  time_t rawtime;
  struct tm* timeinfo;
  time(&rawtime);
  timeinfo = localtime(&rawtime);
  string tempTime = asctime(timeinfo);
  for (int i = 0; i < tempTime.length(); i++) {
    if (tempTime[i] == ' ')
      tempTime[i] = '_';
  }
  string myTime = tempTime.substr(0, tempTime.length() - 1);
  return myTime;
}

string
Util::getTimeString()
{
  static const int max = 20; // max length of time string
  char time_str[max];
  time_t timep;
  time(&timep);
  strftime(time_str, max, "%Y%m%d %H:%M:%S\t", gmtime(&timep));
  return string(time_str);
}

//is ostream.write() ok to update to disk at once? all add ofstream.flush()?
//http://bookug.cc/rwbuffer
//BETTER: add a sync function in Util to support FILE*, fd, and fstream
void
Util::Csync(FILE* _fp)
{
  //NOTICE: fclose will also do fflush() operation, but not others
  if (_fp == NULL) {
    return;
  }
  //this will update the buffer from user mode to kernel mode
  fflush(_fp);
  //change to Unix fd and use fsync to sync to disk: fileno(stdin)=0
  int fd = fileno(_fp);
  fsync(fd);
  //FILE * fp = fdopen (1, "w+");   //file descriptor to file pointer
  //NOTICE: disk scheduler also has a small buffer, but there is no matter even if the power is off
  //(UPS for each server to enable the synchronization between scheduler and disk)
}

string
Util::node2string(const char* _raw_str)
{
  string _output;
  unsigned _first_quote = 0;
  unsigned _last_quote = 0;
  bool _has_quote = false;
  for (unsigned i = 0; _raw_str[i] != '\0'; i++) {
    if (_raw_str[i] == '\"') {
      if (!_has_quote) {
        _first_quote = i;
        _last_quote = i;
        _has_quote = true;
      } else {
        _last_quote = i;
      }
    }
  }
  if (_first_quote == _last_quote) {
    _output += _raw_str;
    return _output;
  }
  for (unsigned i = 0; i <= _first_quote; i++) {
    _output += _raw_str[i];
  }
  for (unsigned i = _first_quote + 1; i < _last_quote; i++) {
    switch (_raw_str[i]) {
    case '\n':
      _output += "\\n";
      break;
    case '\r':
      _output += "\\r";
      break;
    case '\t':
      _output += "\\t";
      break;
    case '\"':
      _output += "\\\"";
      break;
    case '\\':
      _output += "\\\\";
      break;
    default:
      _output += _raw_str[i];
    }
  }
  for (unsigned i = _last_quote; _raw_str[i] != 0; i++) {
    _output += _raw_str[i];
  }
  return _output;
}

//TODO: change these compare functions from int to unsigned, but take care of the returned values
int
Util::_spo_cmp(const void* _a, const void* _b)
{
  int** _p_a = (int**)_a;
  int** _p_b = (int**)_b;

  int _sub_id_a = (*_p_a)[0];
  int _sub_id_b = (*_p_b)[0];
  if (_sub_id_a != _sub_id_b) {
    return _sub_id_a - _sub_id_b;
  }

  int _pre_id_a = (*_p_a)[1];
  int _pre_id_b = (*_p_b)[1];
  if (_pre_id_a != _pre_id_b) {
    return _pre_id_a - _pre_id_b;
  }

  int _obj_id_a = (*_p_a)[2];
  int _obj_id_b = (*_p_b)[2];
  if (_obj_id_a != _obj_id_b) {
    return _obj_id_a - _obj_id_b;
  }

  return 0;
}

bool
Util::parallel_spo_cmp(int* _a, int* _b)
{
  int _sub_id_a = _a[0];
  int _sub_id_b = _b[0];
  if (_sub_id_a != _sub_id_b) {
    return _sub_id_a < _sub_id_b;
  }

  int _pre_id_a = _a[1];
  int _pre_id_b = _b[1];
  if (_pre_id_a != _pre_id_b) {
    return _pre_id_a < _pre_id_b;
  }

  int _obj_id_a = _a[2];
  int _obj_id_b = _b[2];
  if (_obj_id_a != _obj_id_b) {
    return _obj_id_a < _obj_id_b;
  }

  return 0;
}

int
Util::_ops_cmp(const void* _a, const void* _b)
{
  int** _p_a = (int**)_a;
  int** _p_b = (int**)_b;

  int _obj_id_a = (*_p_a)[2];
  int _obj_id_b = (*_p_b)[2];
  if (_obj_id_a != _obj_id_b) {
    return _obj_id_a - _obj_id_b;
  }

  int _pre_id_a = (*_p_a)[1];
  int _pre_id_b = (*_p_b)[1];
  if (_pre_id_a != _pre_id_b) {
    return _pre_id_a - _pre_id_b;
  }

  int _sub_id_a = (*_p_a)[0];
  int _sub_id_b = (*_p_b)[0];
  if (_sub_id_a != _sub_id_b) {
    return _sub_id_a - _sub_id_b;
  }

  return 0;
}

bool
Util::parallel_ops_cmp(int* _a, int* _b)
{
  int _obj_id_a = _a[2];
  int _obj_id_b = _b[2];
  if (_obj_id_a != _obj_id_b) {
    return _obj_id_a < _obj_id_b;
  }

  int _pre_id_a = _a[1];
  int _pre_id_b = _b[1];
  if (_pre_id_a != _pre_id_b) {
    return _pre_id_a < _pre_id_b;
  }

  int _sub_id_a = _a[0];
  int _sub_id_b = _b[0];
  if (_sub_id_a != _sub_id_b) {
    return _sub_id_a < _sub_id_b;
  }

  return 0;
}

int
Util::_pso_cmp(const void* _a, const void* _b)
{
  int** _p_a = (int**)_a;
  int** _p_b = (int**)_b;

  int _pre_id_a = (*_p_a)[1];
  int _pre_id_b = (*_p_b)[1];
  if (_pre_id_a != _pre_id_b) {
    return _pre_id_a - _pre_id_b;
  }

  int _sub_id_a = (*_p_a)[0];
  int _sub_id_b = (*_p_b)[0];
  if (_sub_id_a != _sub_id_b) {
    return _sub_id_a - _sub_id_b;
  }

  int _obj_id_a = (*_p_a)[2];
  int _obj_id_b = (*_p_b)[2];
  if (_obj_id_a != _obj_id_b) {
    return _obj_id_a - _obj_id_b;
  }

  return 0;
}

bool
Util::parallel_pso_cmp(int* _a, int* _b)
{
  int _pre_id_a = _a[1];
  int _pre_id_b = _b[1];
  if (_pre_id_a != _pre_id_b) {
    return _pre_id_a < _pre_id_b;
  }

  int _sub_id_a = _a[0];
  int _sub_id_b = _b[0];
  if (_sub_id_a != _sub_id_b) {
    return _sub_id_a < _sub_id_b;
  }

  int _obj_id_a = _a[2];
  int _obj_id_b = _b[2];
  if (_obj_id_a != _obj_id_b) {
    return _obj_id_a < _obj_id_b;
  }

  return 0;
}

bool
Util::spo_cmp_idtuple(const ID_TUPLE& a, const ID_TUPLE& b)
{
  if (a.subid != b.subid) {
    return a.subid < b.subid;
  }

  if (a.preid != b.preid) {
    return a.preid < b.preid;
  }

  if (a.objid != b.objid) {
    return a.objid < b.objid;
  }

  //all are equal, no need to sort this two
  return false;
}

bool
Util::ops_cmp_idtuple(const ID_TUPLE& a, const ID_TUPLE& b)
{
  if (a.objid != b.objid) {
    return a.objid < b.objid;
  }

  if (a.preid != b.preid) {
    return a.preid < b.preid;
  }

  if (a.subid != b.subid) {
    return a.subid < b.subid;
  }

  //all are equal, no need to sort this two
  return false;
}

bool
Util::pso_cmp_idtuple(const ID_TUPLE& a, const ID_TUPLE& b)
{
  if (a.preid != b.preid) {
    return a.preid < b.preid;
  }

  if (a.subid != b.subid) {
    return a.subid < b.subid;
  }

  if (a.objid != b.objid) {
    return a.objid < b.objid;
  }

  //all are equal, no need to sort this two
  return false;
}

bool
Util::equal(const ID_TUPLE& a, const ID_TUPLE& b)
{
  if (a.subid == b.subid && a.preid == b.preid && a.objid == b.objid) {
    return true;
  }
  return false;
}

void
Util::empty_file(const char* _fname)
{
  FILE* fp;
  //NOTICE: if exist, then overwrite and create a empty file
  fp = fopen(_fname, "w");
  if (fp == NULL) {
    printf("do empty file %s failed\n", _fname);
  } else {
    fclose(fp);
  }
}

//require that _base>=1
unsigned
ceiling(unsigned _val, unsigned _base)
{
  //WARN: we donot check overflow here
  return (_val + _base - 1) / _base * _base;
}

long
Util::read_backup_time()
{
  ifstream in;
  in.open(Util::profile.c_str(), ios::in);
  if (!in) {
    return Util::gserver_backup_time;
  }
  int buf_size = 512;
  char lbuf[buf_size];
  while (!in.eof()) {
    in.getline(lbuf, buf_size);
    regex_t reg;
    char pattern[] = "^\\s*BackupTime\\s*=\\s*((0|1)[0-9]|2[0-3])[0-5][0-9]\\s*(\\s#.*)?$";
    regcomp(&reg, pattern, REG_EXTENDED | REG_NOSUB);
    regmatch_t pm[1];
    int status = regexec(&reg, lbuf, 1, pm, 0);
    regfree(&reg);
    if (status == REG_NOMATCH) {
      continue;
    } else if (status != 0) {
      in.close();
      return Util::gserver_backup_time;
    }
    for (int i = 11; i < buf_size && lbuf[i]; i++) {
      if (lbuf[i] >= '0' && lbuf[i] <= '9') {
        in.close();
        return 36000 * (lbuf[i] - '0') + 3600 * (lbuf[i + 1] - '0')
               + 600 * (lbuf[i + 2] - '0') + 60 * (lbuf[i + 3] - '0');
      }
    }
  }
  in.close();
  return Util::gserver_backup_time;
}